Associations of a National Tax on Non-Essential High Calorie Foods with Changes in Consumer Prices

Abstract

Several governments are considering taxes on non-essential energy-dense, high calorie foods (NEDF) to increase their prices and thereby encourage better diet and health. Alongside a tax on sugary drinks in January 2014, Mexico implemented such a tax: an 8 percent ad-valorem tax on NEDF, defined as those with energy density equal or larger than 275kcal/100g. We study the changes in the prices of taxed and tax-exempt foods following this tax both on average and by tax-eligible foods across store types and cities, using monthly price data between 2012 and 2016. We compare within-product price changes before and after the tax adjusting for product fixed effects, seasonality, and trends, and find that prices of taxed foods increased by 4.8 % on average, but differentially across foods. Prices of candies, cookies and packaged pastries increased by eight or more percent post-tax (vs pre-tax); prices of cakes, and savory snacks increased by less. Prices of fresh pastry and ready-to-eat cereal increased, but only in 2014. Prices of chocolate and pizza did not increase after the tax. For tax-exempt foods, no significant price changes were observed. Variability in price changes for taxed foods were observed by cities as well as by stores: increases were larger in supermarkets compared to smaller grocery stores on average and for most foods. Differences in how prices changed across foods, cities and stores have implications for who is likely to be affected by the tax and how tax effects on diet may vary due to the differential tax pass-through in addition to a heterogenous demand response to changed prices.

Keywords: Mexico; energy-dense foods; prices; taxes.

Figure 1:

Changes in Real Prices of Taxed and Tax-Exempt Foods (2012–2016 vs 2013q4). Notes: Weighted monthly price for taxed and tax-exempt foods was calculated as the mean over a one-month period weighted by the 2010 budget share of the product category that each product for which price is observed belongs to; prices were expressed per kilogram, were log-transformed, de-trended and expressed relative to an average price in 2013q4. Prices were expressed in 2011 Mexican Pesos. Red vertical line marks tax implementation in December 2013.

Figure 2:

Adjusted Changes in Real Prices of Taxed and Tax-Exempt Foods Relative to 2013q4. Notes: a) Quarterly changes in real prices of taxed and tax-exempt foods in % relative to 2013q4 were adjusted for time-invariant product characteristics, linear monthly trend, and product-category, city-level and point-of-sale specific trend. Outcome variable is a log-transformed real price per product in 2011 Mexican pesos. b) Nominal prices were deflated using the January 2011 CPI into real prices in Mexican Pesos, and were expressed per product per 1 kilogram (or liter) per month. c) The percent change in price was calculated by multiplying the estimate on post-tax quarterly indicator variable by 100. 95% confidence intervals were reported. Vertical red line marks the last quarter prior to the tax implementation (2013q4).

Figure 3:

Changes in Prices by Food Categories (2014–2016 vs 2012–2013). Note: a) Changes in real prices of taxed foods in % (2014–2016 vs 2012–2013) were adjusted for time-invariant product characteristics, linear monthly trend, seasonality, and product-category and city-level specific trend. Outcome variable is a log-transformed real price per product in 2011 Mexican pesos. b) Nominal prices were deflated using the January 2011 CPI into real prices in Mexican Pesos, and were expressed per product per 1 kilogram (or liter) per month. c) The percent change in price was calculated by multiplying the estimate on post-tax quarterly indicator variable by 100. d) 95% robust confidence intervals are reported. Ad-valorem NEDF tax marked at 8%.

Figure 4:

Changes in Prices by Store Types. Note: 95% robust confidence intervals are reported. Solid and dashed line describes trajectories for prices collected in supermarkets and mini-markets, respectively. (A) Changes in real prices of taxed foods in % (2014–2016 vs 2012–2013) were adjusted for time-invariant product characteristics, linear monthly trend, seasonality, and point-of-sale, product-category and city-level specific trend. The percent change in price was calculated by multiplying the estimate on post-tax indicator variable by 100. Ad-valorem NEDF tax marked at 8%. (B) Quarterly changes in real prices of taxed foods in supermarkets and mini-markets or convenience stores in % relative to 2013q4 were adjusted for time-invariant product characteristics, and city-level and product category specific trend. The percent change in price was calculated by multiplying the estimate on post-tax quarterly indicator variable by 100. Vertical red line marks 2013q4.

Figure 5:

% Changes in Prices of Taxed Foods Across Cities (2014–2016 vs 2012–2013). Note: Changes in real prices of taxed foods in % (2014–2016 vs 2012–2013) were adjusted for time-invariant product characteristics, linear monthly trend, seasonality, and point-of-sale, product-category, and city-level specific trend. Three cities where prices decreased (Tepatitlan, Tehuantepec, Tapachula) are not presented here. The percent change in price was calculated by multiplying the estimate on post-tax indicator variable by 100. 95% robust confidence intervals are reported.